Foto del docente

Daniela Cocchi

Full Professor

Department of Statistical Sciences "Paolo Fortunati"

Academic discipline: SECS-S/01 Statistics

Research

Keywords: environmental statistics entropy sampling desing bayesian statistics Hierarchical models small area estimation record linkage finite population inference

Methods for finite population sampling.

Statistical models for environmental problems

Methods for finite population sampling.

1. Bayesian solutions for hierarchical superpopulation models and small area inference.

2. Quasi-linear Bayesian approximations in finite populations, in the presence of asymmetry and kurtosis: solution of non standard models, with strong asymmetry and variable data quality.

3. Obtaining a part of Census information by means of a survey: Assessment of alternative proposals for the sampling plan, related both to the population definition and the strategy planning. The strategy  is composed by a sampling plan and an estimator. Auxiliary information for municipalities is the instrument for constructing the sampling plan.

 

Statistical models for environmental problems.

1. Development of hierarchical models and methods for air quality indices.

2. Hierarchical models for Ensemble Weather Forecasting, which are characterized by two main difficulties: first the number of replicated scenarios is limited by the computing costs, second, the forecasting error is not evaluated by means of observed data. The forecasting error is modelled by means of observed data for temperature.

3. Proposal of robust models for evaluating classification errors in measurement systems capability models under a two-component error model